Revisiting the anasynthetic spiral

Grammaticalization is nowadays often seen primarily as a kind of semantic-pragmatic change, but in the 19th century it was more typically seen in a holistic typological pespective: The idea was that synthetic languages develop from analytic languages, and that they may become analytic again. This kind of development is indeed occasionally observed in entire languages, as in the Romance languages and in Later Egyptian, but it is quite unclear whether such holistic changes are at all common. Similarly, there seems to be no good evidence that changes from agglutinative patterns to isolating patterns go through an intermediate flective or fusional stage. By contrast, there is abundant evidence for the old observation that older tightly bound constructions often get competition from new constructions based on content items, which may eventually replace the older patterns (I call this kind of process anasynthesis). Such anasynthetic changes are driven by inflationary processes that can be observed elsewhere in language and culture, not by therapeutic motivations.